2.1.

Skin Samples

Skin samples were obtained from healthy volunteers (phototypes II to III) after written consent. Each of eight volunteers gave two skin biopsies, $6\mathrm{mm}$ in diameter, one from the neck and one from the buttock. Samples from the neck were included in this study in order to be able to distinguish effects of extended UV irradiation due to sun exposure from fs-NIR-irradiated skin. After excision, the skin samples were cut in half.

2.2.

Irradiation

Before skin sample excision, the sensitivity of the volunteers to UV-irradiation was quantified. One minimal erythema dose (MED) was evaluated by visual assessment of six spots irradiated using increasing radiation power of the UV source. The erythema was defined according to COLIPA (The European Cosmetic Toiletry and Perfumery Association) as the first perceptible, clearly defined, unambiguous reddening of the skin $24\mathrm{h}$ after irradiation. UV- and fs-irradiations were performed on freshly excised samples (maximum $1\mathrm{h}$ after excision; see Fig. 1 ).

Fig. 1

Sample preparation and irradiation procedure. (a) Biopsy is divided and irradiated using fs-irradiation (left) or UV-irradiation (right). (b) Fs-irradiation regime 1to $3=10$ horizontal scans around the basal membrane; fs-irradiation regime $4=150$ horizontal scans from the skin surface to the dermis.

For the UV-irradiation, the whole surface of one half of the sample was irradiated using a solar simulator (SU 5000, mut GmbH, Hamburg, Germany) at a UV dose according to 1.5 MED. The UV-irradiation took about $15\min$ . After the UV-irradiation, the samples were frozen in liquid nitrogen for further use. A spectrum of the solar simulator emission is shown in Fig. 2 .

Fig. 2

Emission spectrum of the solar simulator SU 5000.

The other half of the skin sample was irradiated using an ultrashort pulsed, focused laser beam of a Mai-Tai laser (Spectra Physics, Darmstadt, Germany) at a wavelength of $800\mathrm{nm}$ , using a repetition rate of $80\mathrm{MHz}$ and a pulse width at sample layer of about $150\mathrm{fs}$ . For scanning and focusing purposes, a MPLSM (DermaInspect, Jenlab, Jena, Germany) equipped with a Zeiss Plan/Neofluar $40\times$ NA 1.3 objective lens (oil immersion) was used. The DermaInspect device contains a laser emitting NIR in the range between 750 and $850\mathrm{nm}$ according to the specification by the manufacturer. The focus of the laser beam was scanned at various depths within the epidermis and papillary dermis of the excised half of the skin sample. Each horizontal scan was $200\times 200\text{microns}$ wide and took $32\mathrm{s}$ . Automatic horizontal scanning (parallel to the skin surface) using a fixed step width in the depth of the microscope focus was performed. Four laser irradiation regimes were chosen (Fig. 1):

Peak fluence (or peak radiant fluence) is the time integral of the spherical peak irradiance; it combines the two main paramters influencing dosimetry and hence photochemical effects, in our case, the irradiance and the time. The wavelength and the irradiation time were kept constant during experiments in order to investigate the effects of one parameter, the peak irradiance.

The first three regimes were applied around the basal membrane in order to make sure that the most sensitive cells for risk evaluation, the basal layer cells, will receive the laser light. Three different laser powers (15, 30, and $60\mathrm{mW}$ ) were applied, because with tissue imaging, these laser powers had been shown to cover the possible working range when imaging horizontal scans at basal membrane depth. Laser power was measured using a power meter (Ophir Laserstar) with a coverglass and immersion oil in front of the detector. In this depth, at $15\mathrm{mW}$ , imaging starts to be possible. At $30\mathrm{mW}$ , images can be taken with good contrast. At $60\mathrm{mW}$ , first damaging occurs during scanning (fluorescent scars, as described elsewhere⁴). The fourth regime covered the whole depth from the surface to the papillary dermis using linearly increasing laser power (Fig. 1). The most suitable laser power to achieve best imaging results was used in order to simulate tissue imaging as one possible in vivo application.

The four irradiation regimes were applied side by side on the second half of the original skin sample, producing irradiated areas of about $200\times 200\mu \mathrm{m}$ . Between irradiated areas, a space of $400\text{-}\mu \mathrm{m}$ unirradiated tissue was left in order to avoid cross talk between irradiated areas due to light scattering in the skin. At each side of the row of the four irradiated areas, the skin was marked with a black spot in order to retrieve the tiny irradiated areas for the subsequent CPD analysis. The whole fs-irradiation took about $15\text{to}20\min$ ; samples were frozen in liquid nitrogen at $-196°\mathrm{C}$ after fs-irradiation for further analysis.

2.3.

Cyclobutane-Pyrimidin-Dimer Antibodies and Staining Protocol

Sections ( $7\mu \mathrm{m}$ thick) were cut from frozen tissue using a cryostat (Microm, Walldorf, Germany) and were sequentially fixed after slide preparation for $5\min$ in methanol and acetone at $-20°\mathrm{C}$ . After fixation, the slides were stored at $-20°\mathrm{C}$ . For CPD staining, the samples were rehydrated in TKT 100 ( $50\mathrm{mM}$ Tris pH 7.2; $1\mathrm{M}$ KCl, 0.3% Triton X-100) for $60\min$ at room temperature, washed once in phosphate buffered saline (PBS) for $5\min$ and two times for $5\min$ in $2\times$ SSC ( $300\mathrm{mM}$ NaCl, $30\mathrm{mM}$ Na citrate). The sections were denatured in $0.1\mathrm{N}$ NaOH/70% EtOH for $3\min$ ; dehydrated for $1\min$ each in 70%, 90%, and 100% EtOH; air dried; and incubated with Proteinase K ( $10\mu \mathrm{g}$ per ml; Sigma, Taufkirchen, Germany) at $37°\mathrm{C}$ for $10\min$ . After three washes for $5\min$ each in PBS, they were incubated with 5% goat serum (Dako, Hamburg, Germany) for $30\min$ at room temperature, rewashed three times for $5\min$ in PBS, and incubated overnight at $4°\mathrm{C}$ with monoclonal antibody specific for CPD (Kamiya Biochemical, Seattle, Washington) diluted 1:1000 in PBS. Sections were then washed three times for $5\min$ each in PBS and incubated with goat anti-mouse IgG conjugated with fluorescein isothiocyanate (FITC; Dianova, Hamburg, Germany) diluted 1:100 in PBS. After three additional $5\text{-}\min$ washes in PBS, slides were mounted with “antifade” (2.3% DABCO; Sigma, Taufkirchen, Germany) in 90% glycerol in $20\mathrm{mM}$ Tris, pH 8.0 and covered. Sections were visualized using a CCD camera (Kappa, Gleichen, Germany) coupled to a fluorescence microscope (Leica DM, Leica, Wetzlar, Germany). Fluorescence intensity was quantified using a digital imaging system (OPTIMAS, Bothell, Washington). Under the premise that the images of the different samples were taken under the same camera conditions and a careful handling of the probes to minimize the influence of photobleaching, the fluorescence intensity is a reliable measure for the amount of induced CPDs (e.g., Refs. 26, 27). Comparing the data of laser-induced CPDs with those induced by 1.5 MED solar UV radiation (UVR), the method can be used to establish the possibility for estimation of skin cancer risks according to current models.

4.1.

UV- and Fs-NIR-Irradiation Effects in Human Skin

In the unirradiated areas of fs-exposed samples and untreated buttock samples, no CPD-related fluorescence could be observed [Figs. 4a and 4c]. In contrast, unirradiated areas of fs-exposed neck samples showed already a high amount of CPDs, corresponding to 48% of CPDs induced by a single 1.5 MED solar-simulated UV-radiation (UVR). Since the neck is a body site that is naturally heavily sun exposed, it is assumed that the detected CPDs were induced by solar irradiation. Although CPDs can be repaired by epidermal cells very efficiently,^{28, 29, 30, 31, 32} it has been observed that even $10\text{days}$ after irradiation, about 2% of epidermal cells still contain CPDs.³³ In all the investigated neck samples here, heavily damaged cells in the basal layer of the epidermis can be observed (Fig. 3a]. These CPD-retaining basal cells (CRBCs) occur in human skin depending on the amount of sun exposure. Some evidence exists that they may represent epidermal stem cells,³⁴ which are thought to be the most important targets for cutaneous cell aberrations, possibly leading to skin carcinogenesis.³⁵

In the fs-irradiated buttock samples of all volunteers, a region with CPD-related fluorescence of cell nuclei was observed that can be attributed to the $60\text{-}\mathrm{mW}$ fs-NIR-irradiation scheme [see Fig. 5a, fs-buttock, area 1]. This region has a maximum area of $2\times {10}^{-3}{\mathrm{mm}}^2$ [Fig. 4c]. Additionally, a second region could be found in three out of eight volunteers [see Fig. 5b, fs-buttock, area 2]. This region can be attributed to the area irradiated by the scheme using the z-scan with increasing laser power (regime 4). These second CPD-containing regions are small, representing $2.5\times {10}^{-4}{\mathrm{mm}}^2$ [Fig. 4d], respectively. Obviously, after $60\text{-}\mathrm{mW}$ fs-irradiation with a step width of $5\mu \mathrm{m}$ or after 150 horizontal scans with linearly increasing laser power ( $2\mathrm{mW}$ to ca. $35\mathrm{mW}$ ) and a step width of 1 micron, the number of two-and three-photon excitations is sufficient to induce CPDs in deeper cell layers. Taking into account that the microscope focus has a $z$ dimension of about $1.5\mu \mathrm{m}$ , a double exposure occurs in every $1\text{-}\mu \mathrm{m}$ scan-step of two consecutive scans and enhances the formation of CPDs locally. On the other hand, due to skin scattering, deeper tissue layers will receive less power. Sun-simulator-exposed skin, on the other hand, shows significant CPD formation in all viable cell layers of the epidermis and dermis, and in the basal layer not limited to a restricted area.

4.2.

Assessment of the Biological Effectiveness of UV Damage by Fs-NIR-Laser Irradiation

For a better assessment of the damage induced by fs-NIR-laser irradiation, one has to compare the amount of CPD formation with the natural amount of DNA damage after sun exposure and evaluate the amount of remaining CPD based on data from the literature. The outermost risk of UV irradiation can lead to three types of skin cancer, basal cell carcinomas (BCC), squamous cell carcinomas (SCC), and malignant melanoma (MM).^{36, 37} The first two types have a relatively low lethality. On the other hand, late malignant melanoma stages lead to the death of the patient in a number of cases, if the melanoma is not removed in time. The connection between UV irradiation and the appearance of BCC or SCC is well proven epidemiologically. With MM, the connection is likely, but not equally well understood. Using animal investigations for nonmelanoma tumors, an action spectrum has been defined by de Gruijl and van der Leun.⁴¹

In the UVB and UVA range, this action spectrum follows the erythema action spectrum of the Commission Internationale de L’Eclairage (International Commission on Illumination, CIE; see dotted line, Fig. 6 ). Due to the absorption of shorter wavelengths in the skin the radiation is less effective in the UVC range, reaching a maximum at a wavelength of $290\text{to}310\mathrm{nm}$ , as these data include the protecting effect of the upper skin layers in the estimation of damage induction in deeper epidermis layers (e.g., basal cells).

Fig. 6

Action spectrum for the development of CPDs (▲) in the epidermis compared to the action spectrum for skin cancer induction (●). For comparison, the erythema action spectrum of the Commission Internationale de L’Eclairage (International Commission on Illumination, CIE) is included (dotted line). Modified from Ref. 24; additional references can be found there.

Since the primary laser irradiation of the MPLSM used has a typical wavelength range of $750\text{to}850\mathrm{nm}$ , acute direct DNA damage by NIR radiation can be excluded. Hence, a real consideration of DNA damage has to take into account excitation of energy states usually accessible only by UV absorption but now generated by multiphoton excitation processes in viable skin layers. Further, the natural DNA repair efficiency after a single exposure by the fs-NIR-laser has been estimated.

One has to bear in mind, however, that the action spectra for BCC and SCC, as shown in Fig. 6, could be different in the fs-irradiation situation, because they are related to the effect of irradiation dose at the skin surface. For the mutation induction in mammalian cells not protected by cornified skin layers, an action spectrum for cells in culture was published previously.³⁸ This can also be applied mainly to deeper skin layers (e.g., basal layer) because they resemble a kind of 3-D cell culture domain.

In comparison to the BCC/SCC action spectrum, a much higher effectiveness can be observed at $250\text{to}280\mathrm{nm}$ in nonprotected cells. This is important, since in addition to the main $425\text{to}325\mathrm{nm}$ , two-photon excitation, there could be a nonnegligible amount of short-wavelength UV effects around $267\mathrm{nm}$ being produced in deeper skin layers due to three-photon excitation produced by ultrashort focused laser irradiation.

The fs-NIR-irradiation experiments have proven that the energy density is high enough to lead to CPD formation, but whether these CPDs were induced by a two-or three-photon process remains open at the moment.

At a minimum, the measurement of the amount of CPDs offers a possibility of a very rough quantitative risk estimation at the level of DNA damage. The parallel exposure to 1.5 MED simulated sun radiation can serve as a “biological reference” leading to an exposure of 0.6 and 0.45 MED by fs-NIR-irradiation regimes 3 and 4, respectively. Hence, it is not necessary to measure the local photon dose in deeper skin layers. For BCC and SCC, a risk estimation was performed.³⁹ With a yearly UV dose of 100 MED over $30\text{years}$ at ages from 15 to 45, the risk increases until age 75 by a factor 4. Assuming a linear dose relationship—as is usual in this kind of considerations³⁹—in a rough estimation, one can conclude that one additional exposure by fs-NIR-irradiation regime 3 or 4 increases the risk locally by 0.6% or 0.45%, respectively.

These estimations do not include the fact that only very small, defined areas are irradiated (e.g., in our case, $200\times 200\mu \mathrm{m}=0.04{\mathrm{mm}}^2$ ) by MPLSM investigations compared to a sun exposure of total body parts or the whole body. For example, if the face is sun exposed, its area of about $10\times 10\mathrm{cm}$ $(={10}^4{\mathrm{mm}}^2)$ is by a factor 250,000 larger than that exposed during fs-NIR-irradiation.

One has to keep in mind that the risk estimation is based on standard values for a Caucasian skin with phototype II or III and on limited number of ex vivo experiments with a limited number of possible irradiation parameters. Subjects with phototype I are more sensitive to cancer risk due to UV-irradiation and may be more sensitive to fs-irradiation as well. Since the tissue damage is caused by a nonlinear effect, changing laser parameters may result in more severe tissue damage (see Ref. 40).

If the mechanisms of DNA damage differ from that of UV-irradiation, especially when three or more photons are involved that excite to energy states not be reached by UV-irradiation due to the protection of the upper layer of the skin, these experiments may underestimate the cancer risk. In addition, a specially targeted irradiating of cells that are especially important for tumorgenesis (e.g., CRBCs) can increase the tumor risk over a linear extrapolation of the dose or irradiated area.

For future work, it will be interesting to see which role the DNA repair process may play after fs-NIR-irradiation and whether CRBCs can be induced or remain. Furthermore, it is important to investigate alternative UV-induced DNA lesions like 8-oxo-guanine or, even more important, DNA double-strand breaks, which recently have been shown to be induced by UVA-radiation. Nevertheless, the large amount of CPDs produced in the neck area after natural sun exposure is usually repaired by the DNA polymerase repair system in a very efficient way.

Taking all the findings and considerations together, at our irradiation parameters, the risk increase by an fs-NIR-irradiation of a defined small skin area can be assumed to be reduced to negligible values. Even several fs-NIR-irradiation sessions delivered at different $0.04\text{-}{\mathrm{mm}}^2$ areas will not increase the risk of remaining DNA damage in a measurable way.